FRBRizing Amazon’s Catalog (March 24th, 2008)

While doing research for a book data project, I stumbled on an interesting discovery. One movement that has being gaining steam recently among librarians and others involved with book information is FRBR which among other things seeks to link various editions of the same book together. For example, all of Harry Potter books are currently entered into library and book store catalogs separately and there is no easy way to find a translation of a specific volume. FRBR seeks to fix that by re-thinking catalogs in a hierarchy where the Work (in this case Harry Potter) will be linked to each of its editions and translations (Expressions and Manifestation).

Unfortunately, it is not an easy task to figure out what links to what. There are two current public approaches – OCLC which operates the WorldCat service used computers to automatically try to match up different editions. Their service is called xISBN and is limited to 500 queries a day and non-commercial use. OpenLibrary is doing something similar with an algorithm. The other approach is by using people and this is what LibraryThing is doing with their thingISBN service. People who use LT to catalog their books have an option to specify if specific editions are in fact the same work. That cumulative data is published via their API.

About two weeks ago I accidentally stumbled on a third public service that does something similar.  When Amazon launched their Kindle eBook reader they made lots of titles available as a Kindle eBook. HOWEVER, they did not want to change the ISBN numbers for these titles. So what they did is re-organize their catalog is a way that all editions of the same work now appear to be linked to together including audio, eBook, hard cover, etc. This ability is buried in their API right here and is called RelatedItems:

 The RelatedItems response group returns items related to an item specified in an ItemLookup request. .Related items could be, for example, all of the Unbox episodes in a TV season that are sold separately, or, for example, all of the MP3Download tracks on a MP3 album.

Each item now has a parent “authority title” which lists all of the children which are editions of that work. The authority title corresponds to Work in FRBR and the children are Expressions and Manifestations.

Practically speaking what this means is that book related websites now have another way to figure out how different editions are related to each other. Obviously this service is skewed towards stuff that is actually being sold but it does show some surprising results. For example, a search for one of the Pendragon books shows an extra few editions in Amazon that neither LibraryThing or WorldCat list. On the other hand, searching for Dune shows better data in WorldCat and LT than Amazon.

I am currently working on a possible service that will use Amazon’s data as an supplement to xISBN and thingISBN. Because of legal issues, this will not be a web service but rather code that can be used with people’s own AWS accounts. The service is tentatively called amazingISBN. I am also releasing an experimental web tool that retrieves data from all three services side by side. Due to use restrictions and underlying legal issues, this service is highly experimental.

Advertisements

Comments?

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s